Analysis of protein complexes in wheat amyloplasts reveals functional interactions among starch biosynthetic enzymes.

Protein-protein interactions among enzymes of amylopectin biosynthesis were investigated in developing wheat (Triticum aestivum) endosperm. Physical interactions between starch branching enzymes (SBEs) and starch synthases (SSs) were identified from endosperm amyloplasts during the active phase of starch deposition in the developing grain using immunoprecipitation and cross-linking strategies. Coimmunoprecipitation experiments using peptide-specific antibodies indicate that at least two distinct complexes exist containing SSI, SSIIa, and either of SBEIIa or SBEIIb. Chemical cross linking was used to identify protein complexes containing SBEs and SSs from amyloplast extracts. Separation of extracts by gel filtration chromatography demonstrated the presence of SBE and SS forms in protein complexes of around 260 kD and that SBEII forms may also exist as homodimers. Analysis of cross-linked 260-kD aggregation products from amyloplast lysates by mass spectrometry confirmed SSI, SSIIa, and SBEII forms as components of one or more protein complexes in amyloplasts. In vitro phosphorylation experiments with gamma-(32)P-ATP indicated that SSII and both forms of SBEII are phosphorylated. Treatment of the partially purified 260-kD SS-SBE complexes with alkaline phosphatase caused dissociation of the assembly into the respective monomeric proteins, indicating that formation of SS-SBE complexes is phosphorylation dependent. The 260-kD SS-SBEII protein complexes are formed around 10 to 15 d after pollination and were shown to be catalytically active with respect to both SS and SBE activities. Prior to this developmental stage, SSI, SSII, and SBEII forms were detectable only in monomeric form. High molecular weight forms of SBEII demonstrated a higher affinity for in vitro glucan substrates than monomers. These results provide direct evidence for the existence of protein complexes involved in amylopectin biosynthesis.


Introduction
Starch is produced by the majority of higher plant species inside plastids, and represents a major storage product of many of the seeds and storage organs produced agriculturally and used for human consumption, as well as many important industrial applications. The starch granule is a complex polymeric structure with a hierarchical order, allowing efficient packing of large amounts of glucose into a water-insoluble form, and is composed of two distinct types of glucose polymer; amylose and amylopectin.  (Buléon et al., 1998;Thompson, 2000). Granule formation is driven by both the semi-crystalline properties of amylopectin, as determined by the length of the linear chains of amylopectin and the clustering and frequency of α -(1→6)-branch linkages these activities, as well as other enzymatic steps, including starch degrading activities, is likely to be required to produce the non-random clustered arrangement of glucan chains characteristic of amylopectin.
To date, information on how co-ordination between amylopectin synthesizing enzymes is achieved is sparse. Recent work in wheat endosperm amyloplasts suggests protein phosphorylation is involved in modulating the catalytic activity of some key enzymes (the SBEII class) and their ability to form physical interactions with other starch metabolizing enzymes (Tetlow et al., 2004b). Analysis of transgenic Arabidopsis and potato plants also indicates a role for 14-3-3 proteins in the regulation of SS activity in the formation of assimilatory (transient) starch in leaves, presumably through proteinprotein interactions (Sehnke et al., 2001;Zuk et al., 2005).
The objective of the research described in this communication was to investigate interactions between SSs and SBEs in starch-synthesizing plastids. This article describes the isolation and characterization of protein complexes comprising SSI, SSII, and SBEII from amyloplast extracts of developing wheat endosperm. The data indicate that formation of the SS/SBE protein complexes, and SBEII homodimers, in the developing endosperm occurs from around 10-15 days after pollination (DAP), the major grain filling period, and is phosphorylation dependent. The paper presents the first direct evidence of catalytically active protein complexes involved in amylopectin biosynthesis, and that the indicating an apparent increase in molecular mass, or aggregation state of the enzymes contributing to these measured activities. At later stages of endosperm development, the total eluted SS and SBE activity was divided almost equally between an apparently high molecular weight peak of around 200-300 kDa (termed HMW), as well as the peak originally observed at the earlier stages of endosperm development (6-9 DAP) with an apparently lower molecular mass, corresponding to the size of the monomeric proteins (termed LMW). Analysis of SS and SBE activities by gel filtration chromatography at later stages of endosperm development (beyond 15 DAP) showed essentially the same separation of peaks of SS and SBE activities as observed in Figure 1 at 10-15 DAP (data not shown).
The results described above were obtained with whole cell homogenates. Similar results were also obtained when amyloplast lysates prepared from endosperm were separated by gel filtration chromatography, and fractions assayed for SBE and SS activities (data not shown).
Analysis of the fractions (following gel filtration chromatography of whole cell extracts) by immunoblotting with anti-SBE antibodies ( Figure 1C) showed that both forms of SBEII (SBEIIa and SBEIIb) were responsible for the measured SBE activity in both LMW and HMW fractions, and that SBEI was not expressed at the stages of endosperm development used in these experiments, consistent with previous findings (Morell et al., 1997). Immunoblots of the gel filtration column fractions developed with anti-SSI and anti-SSII antibodies indicated the presence of both these forms at early stages of development (6-9 DPA), which is consistent with the known patterns of SS gene expression in wheat (Li et al., 1999a), and in both the LMW and HMW peaks of activity at later stages of endosperm development ( Figure 1D).

In Vitro Phosphorylation of SS and SBE Forms in Plastids
Previous work has shown that many stromal proteins rapidly become phosphorylated  Figure S1), in agreement with previous work (Tetlow et al., 2004b). However, we found no evidence to suggest that the soluble form of SSI is phosphorylated in amyloplasts. We observed no apparent alterations in the electrophoretic mobility of the different proteins as a result of phosphorylation using the 1D-gradient gel systems.

Coimmunoprecipitation of SS and SBE Forms in Plastids
Amyloplast lysates were used as a source of material for coimmunoprecipitation experiments in order to examine protein-protein interactions and, the possibility that phosphorylation has a role in facilitating such interactions ( Figure 2). Plastid lysates were employed to reduce the cross-reactivity of antibodies with non-specific proteins associated with the use of whole cell extracts. Figure 2 shows the results of experiments using anti-SBEII antibodies as the immunoprecipitation agent. The data in Figure 2A show that each of the peptide-specific anti-SBEII antibodies (and the anti-SS antibodies, Supplementary Fig 1) precipitates only the respective form of SBE and SS from plastid lysates, and that phosphorylation (preincubation with ATP) or dephosphorylation (pre-incubation with APase) does not affect the ability of any of the antibodies to bind to the respective forms of SBE and SS.
Experiments with stroma prepared from amyloplasts isolated from endosperm at 6-9 DAP showed no co-immunoprecipitation of other starch synthesizing enzymes as judged by immunoblotting, regardless of the pre-treatment conditions ( Figure 2A). However, at 10-15 DAP, when higher molecular mass/aggregation states of SSs and SBEs were observed in the gel filtration experiments (Figure 1), both SBEII antibodies were able to coimmunoprecipitate SSI and SSII, indicating potential protein-protein interactions between each of the SBEII forms and SSI and SSII. Pre-incubation with plastid lysates with glucan-degrading enzymes (amyloglucosidase and α -amylase) did not prevent coimmunoprecipitation of SS and SBE isoforms, which indicates that their association is due to specific protein-protein interactions, and not a result of SSs and SBEs binding to a common glucan chain.
Significantly, the SS/SBE interactions observed at 10-15 DAP of endosperm development in Figure 2B were not observed when plastid lysates were pre-incubated with APase (which causes non-specific dephosphorylation), suggesting that the SS/SBE interactions observed in wheat endosperm are developmentally regulated, and also phosphorylation-dependent. Lysates of late stage endosperm amyloplasts which were not incubated with ATP showed the same coimmunoprecipitation phenomena as those incubated with ATP (data not shown) suggesting that the phosphorylation state of the enzymes is not markedly altered by exogenous ATP (given at 1 mM). Reciprocal experiments demonstrate that both forms of SBEII are co-precipitated by antibodies to SSI and SSII. Figure 2C illustrates the co-immunoprecipitation of SBEIIb in a phosphorylation-dependent manner. Immunoblots of proteins immunoprecipitated with anti-SBEII antisera were also developed with other available T. aestivum-specific antibodies; anti-plastidial starch phosphorylase, anti-plastidial 1,4-α-D-glucan:1,4-α-Dglucan, 4-α-D-glucanotransferase (disproportionating enzyme, D-enzyme) and antiisoamylase. None showed coimmunoprecipitation with anti-SBEII and anti-SS antibodies, at the early tissue developmental stages investigated, or as a function of the phosphorylation status of the stromal preparations (data not shown).

Linking
Stromal proteins from amyloplasts isolated from endosperm at 10-15 DAP were separated by gel filtration chromatography. Eluted column fractions were then immediately incubated with the homobifunctional cross-linking reagent bis (sulfosuccinimidyl) suberate (BS 3 ), and the proteins separated by SDS-PAGE, electroblotted and developed with various anti-SBE and anti-SS antisera. Figure 3A shows the electroblotted, cross-linked proteins from column fractions containing the two major peaks of SS and SBE activities, corresponding to the HMW fractions (containing putative protein complexes) and the LMW fractions (containing monomeric forms of SS and SBE, see Figure 1) cross-linked products were obtained when amyloplast lysates were incubated with BS 3 prior to separation of proteins by gel filtration chromatography (data not shown). Figure   3A shows that when proteins separated by gel filtration chromatography are incubated with APase prior to cross-linking with BS 3 (eluted column fractions were pre-treated with APase prior to the addition of BS 3 ), the protein complexes in the HMW fraction dissociate into monomers, and no aggregated SS or SBE products could be detected.
The cross-linked polypeptides which cross-reacted with the various antisera described above, and shown in Figure 3A, were in-gel digested with trypsin (from corresponding silver-stained SDS-gels), and some of the resulting peptides sequenced using quadrupoleorthogonal-acceleration-time of flight mass spectrometry (Q-TOF-MS). The MS survey acquisition data obtained from single representative analyses are shown in Figure 3B. The sequence data in Figure 3B(1) shows that the amyloplast proteins present in the 260 kDa cross-linked complex(es) were SSI, SSIIa and forms of SBEII (the close sequence homology between SBEIIa and SBEIIb means that the two forms cannot be distinguished on the basis of the peptide sequences acquired by the mass spectrometer), whereas only peptides from SBEII forms could be detected in the cross-linked complexes of approximately 180 kDa ( Figure 3B). Figure 3C demonstrates that SSI protein could be detected in association with SSII following immunoprecipitation with a monospecific SSII antibody (Fig S1), confirming that SSI and SSII can co-exist in the same complex.
The proteins from the LMW fraction which cross-reacted with the various antisera shown in Figure 3A were also in-gel digested with trypsin, and the peptides analyzed by Q-TOF-MS; these analyses confirmed that each of the antibodies recognized the respective monomeric protein (data not shown).
but these were of low abundance, and no measurable MS spectra could be obtained from them.
Washed starch granules were also incubated with BS 3 to determine whether any of the granule-associated proteins formed aggregates/complexes. BS 3 is a low molecular weight water soluble cross-linking reagent (mass of 368.4) which would be predicted to penetrate the starch granule structure. Nevertheless, we found no evidence to suggest that any of the granule-associated proteins were present as aggregation products (determined by immunoblotting and silver-stained SDS-gels), even when starch granules were partially digested by α -amylase prior to treatment with BS 3 in order to increase the accessibility of the cross-linker to the granule-associated proteins (data not shown).

Immunoprecipitation of SS Activity with Anti-SBEII Antibodies
In parallel to the previously described analysis of protein pellets after immunoprecipitation, the supernatants were also assayed for residual SS activity. The data presented in Figure 4A show that soluble SS activity is immunoprecipitated by both of the SBEII antisera only in the HMW fractions. In total, both of the anti-SBEII antibodies precipitated approximately 90% of the measurable SS activity in the HMW fraction, each accounting for approximately equal amounts of the soluble SS activity.
Analysis of the pellets following immunoprecipitation of the HMW protein fractions by anti-SBEII antisera indicated the presence of SSI and SSIIa ( Figure 4B). However, addition of anti-SBEIIa or anti-SBEIIb antisera to LMW fractions caused no reduction in soluble SS activity in the supernatant, and only the respective forms of SBEII could be detected in the pellets following immunoblot analysis, indicating no interaction with SS forms ( Figure 4B). Similar results were obtained in an experiment in which both SBEII antibodies were added to the HMW and LMW fractions; approximately 90% of the measurable SS activity in the HMW fraction was precipitated (data not shown).

Analysis of the Glucan-Binding Properties of LMW and HMW SBE Activities
Affinity gel electrophoresis was employed to quantify the dissociation constant ( (Table 1). In contrast to the behaviour of the SBEII forms in the gels, control proteins (bovine serum albumin, and molecular mass standards) showed no change in migration, irrespective of the glucan concentration in the gel (data not shown). The behaviour of the control proteins indicates that the reduced mobility observed with the different SBEII forms in the affinity gels is a result of their specific affinity for the glucan provided, rather than being caused by a dense polyacrylamide/glucan matrix in the gel. Surprisingly, in gels containing no glucan substrate, the mobility of the different proteins tested was the same in LMW and HMW fractions. However, the K d values derived for both isoforms of SBEII in HMW fractions, run on native gels containing increasing concentrations of corn starch, were significantly lower (P < 0.001) than the respective SBEII isoforms from LMW fractions. Table 1 shows that the HMW-forms of SBEII have a 2-fold higher affinity (as measured by the 1/K d values) for the glucan substrates tested than their respective monomeric forms. The kinetic parameters determined for the monomeric forms of SBEII in wheat endosperm are similar to those determined for other branching enzymes (Matsumoto et al., 1990). The results indicate only small differences between monomeric forms of SBEIIa and SBEIIb in their respective affinities to α -(1→4)-linked glucans.

Discussion
This paper presents evidence for the existence of functional interactions between the SBEII class and the SSI and SSII classes of amylopectin-synthesizing enzymes in amyloplasts from developing wheat endosperm. Changes in elution profile following gel permeation chromatography, demonstration of co-immunoprecipitation of starch synthases with branching enzymes, chemical cross-linking and the loss of protein  Figure 6).
Other forms of SS and SBE are known to be expressed in the endosperm; SSIII, SSIV, and SBEI. These isoforms were not detected when the 260 kDa cross-linked aggregates were analyzed by Q-TOF-MS. However we cannot rule out that SSIII and SSIV may also be complexed since we do not have specific probes for these proteins in wheat.
Evidence is provided in an accompanying paper that SSIII also interacts with other The results show that as much as 40-50 % of the measurable SS catalytic activity in the endosperm is in a high molecular weight form at around 10-15 DAP ( Figure 1) although it should be noted that activities of starch synthases (Mu et al, 1994) and branching enzymes (Smith, 1990) may be underestimated in cell extracts, the latter because of interference by amylolytic activities. In the case of starch synthase activity, the amount recovered from the gel permeation column exceeded the measurable activity applied by as much as 3-fold, suggesting that some inhibitory factor(s) may have been removed during fractionation. Whilst this observation complicates interpretation it at least suggests that loss of activity is unlikely.
The coimmunoprecipitation experiments described in Figure 4 show that the SS/SBE protein complexes of around 260 kDa are made up of catalytically active SSs, since each anti-SBEII antibody was able to immunoprecipitate approximately half of the SS activity in the HMW fractions. This observation therefore implies that most of the SS activity in the HMW fractions is associated with SBEII forms.
In addition to demonstrating their presence in heterocomplexes, evidence has been provided from chemical cross-linking studies that SBEII isoforms also form homo-dimers.
To our knowledge, these experiments in wheat and maize are the first published evidence for homo-dimer formation amongst SBEII forms. SBEIIa and SBEIIb show a high degree of sequence identity, so it must be assumed that the less homologous regions at the amino-and carboxy-terminal ends of the proteins drive dimerization. Analysis of the mobility of HMW and LMW forms of SBEII in gels containing presumptive glucan substrates ( Figure 5) suggests that the former may produce a catalytically functional unit with increased affinity for glucan substrates, compared with the monomeric SBEII forms.
This difference in affinity could represent the activity due to formation of homodimers of either SBEII isoform, or be the result of physical interaction with starch synthases in heteromeric complexes. The HMW and LMW forms of SBEII showed identical mobilities in native gels containing no glucan substrate ( Figure 5), which is more likely to be consistent with the activity of dimers. This is reinforced by the observation that immunodetection of SSI on the same blots did not show the same relative mobility of either form of SBEII (data not shown). The possibility that the differing affinity for starch may have arisen from the formation of a heteromeric complex cannot be ruled out completely since it is possible that the complexes containing SBEII and SS might disassemble during electrophoresis. What is clear is that SBE isoforms in the HMW fraction show markedly different kinetic properties to their monomeric counterparts. Previous genetic and biochemical data are consistent with the existence of protein-protein interactions among amylopectin biosynthetic enzymes, in particular, between SBE and SS forms. For example, the du 1maize mutant which conditions a loss of SSIII function (Gao et al., 1998;Cao et al., 1999) also causes a decrease in SBEIIa activity (Boyer and Preiss, 1981;Cao et al., 2000). In rice endosperm, the ae mutation (causing a loss of SBEIIb activity) also shows a significant (50%) reduction in the activity of soluble SSI (Nishi et al., 2001). Loss of SSIIa in wheat, barley and rice endosperms also causes a reduction in amylopectin synthesis, and abolishes the presence of SSI, SBEIIa and SBEIIb within the starch granules (Yamamori et al., 2000;Morell et al., 2003;Umemoto and Aoki, 2005). All of these genetic observations could be explained by interactions between specific enzymes within a complex. the SSs and SBEs (see above), additional in vitro evidence exists for functional interactions between these enzyme classes. In maize kernel extracts, the activity of SSI was greatly stimulated by the addition of purified SBEI or SBEII (Boyer and Preiss, 1979), and Seo et al. (2002) showed that functional interactions exist between heterologously expressed SBEs from maize, and yeast glycogen synthases, which were proposed to work in a cyclically interdependent fashion, which is consistent with the idea that SSs and SBEs may operate within hetero-protein complexes. Functional assemblies of this kind would presumably improve the efficiency of polymer construction as the product of one reaction becomes a substrate for another within the complex (substrate channelling  Morell et al., 2003;Tetlow et al., 2004b), and the results presented here, indicate that the distinct structure of amylopectin is probably the product of many combinations of interacting enzymes, some of which are components of protein complexes which may be active or inactive at different times. Future studies will focus on determining the distinct glucan products produced by the action of the protein complexes described, and identifying the regulatory proteins involved in assembly and disassembly of the protein complexes.

Plants and Growth Conditions
Spring wheat (Triticum aestivum L. cv. Taho

Protein Extraction from Developing Endosperm
Whole cell extracts were prepared by rapidly homogenising approximately 0.5-0.8 g endosperm (of various individual seed weights corresponding to different stages of development) in 1cm 3 of ice cold rupturing buffer, followed by centrifugation at 13,500xg for 2 min at 4 o C. The resulting supernatant was subjected to ultracentrifugation as described above and immediately loaded onto the size exclusion column (below).

Size Exclusion Chromatography
Amyloplast lysates and whole cell extracts were separated by size exclusion chromatography using a Superdex 200

Enzyme Assays
Sub-cellular marker enzyme assays were performed as previously described (Tetlow et al., 2003). SBE activity was assayed semi-quantitatively using a modification of the phosphorylase a stimulation assay described by Smith (1988 Reactions were terminated by heating the mixture at 90°C for 5 min and loading the mixture onto anion-exchange resin columns (AG ® 1-X8 resin, Bio-Rad Laboratories [Canada] Ltd., Mississauga, Ontario) as described by Jenner et al. (1994). The modified SBE and SS assays were optimized with respect to substrate concentration/glucan primer used, and reactions were all linear with respect to protein concentration and reaction time prior to experimentation.
Amylolytic activity was estimated by measuring the release of 14 C-labelled products from a 14 C-labelled glycogen substrate. The 14 C-labelled glycogen substrate was prepared by incubating 1.5 units of phosphorylase a (product number P-1261, Sigma-Aldrich) in a buffer containing 100 mM sodium citrate (pH 8), 1mM Na 2 -EDTA, 1 mM DTT, 2.5 mM AMP, 4.8 mg cm -3 glycogen (from rabbit liver, type III, Sigma-Aldrich), and 10 mM [U-14 C]-glucose 1-phosphate (3.7-7.4 kBq per preparation) for 2 h at 25°C. The reaction was terminated by heating at 95°C for 5 min and the radio-labelled glucan washed in methanol-KCl as for SBE assays above. The assay for amylolytic activity involved incubating 0.1 cm 3 of washed 14 C-labelled glucan with 0.1 cm 3 protein fractions for 10 min at 25°C. The reaction was terminated by heating to 95°C for 5 min and the remaining into the supernatant were taken as a measure of amylolytic activity and counted using a liquid scintillation counter.

Phosphorylation of Amyloplast Proteins In Vitro
Phosphorylation reactions in which intact amyloplasts were incubated with γ -32 P-ATP were performed as described previously (Tetlow et al., 2004b). Reactions were terminated by lysis in ice-cold rupturing buffer followed by immediate desalting on NAP-10 columns (Amersham Biosciences, Québec, Canada) which had been pre-equilibrated in rupturing buffer with no protease inhibitor cocktail present. The desalted stromal proteins were used in immunoprecipitation experiments (see below). The phosphorylation status of proteins was also determined using ProQ Diamond™ stain (Invitrogen Canada Inc., Burlington, Ontario) in conjunction with APase as a control and following the manufacturer's instructions.

Preparation of Peptides and Antisera
Polyclonal antibodies were raised in rabbits against the synthetic peptides derived from the N-terminal sequences of wheat SBEI (VSAPRDYTMATAEDGV) and wheat SBEIIa (AASPGKVLVPDGESDDLASY) (Rahman et al., 2001), and wheat SBEIIb (AGGPSGEVMIGC). The antigen was prepared by coupling the synthesized peptide to keyhole limpet hemocyanin using the heterobifunctional reagent m-maleimidobenzoyl-Nhydroxysuccinimide ester. Anti-T. aestivum SSI and anti-SSII antisera were prepared as described by Li et al. (1999b), and anti-T. aestivum D-enzyme antisera was prepared as described by Bresolin et al. (2006).

Immunoprecipitation
Immunoprecipitation experiments were performed with samples of amyloplast stroma, and size exclusion column fractions, using methods previously described (Tetlow et al., 2004b). SBE antibodies were added to plastid stroma at concentrations described

Zymogram Analysis and Affinity Electrophoresis
For native zymogram analysis, protein samples were mixed with native gel sample buffer Inc.) for 2 h at 4°C. Gels were washed, incubated for 2 h at 30°C, and stained, as previously described (Nishi et al., 2001). Gels were photographed immediately after staining. The different SBE isoforms from wheat endosperm separated on the native gels were identified as described previously (Tetlow et al., 2004). 100V constant at 25°C in running buffer (25 mM Tris, 192 mM glycine) containing 1 mM DTT. The migration distances of the proteins were measured after immunoblotting and probing for specific enzymes as described above.

Calculation of Dissociation Constants
Affinity electrophoresis was used as a means of measuring protein-glucan interactions, and dissociation constants are calculated from the retardation of the electrophoretic mobility of enzyme/protein by the substrate contained in the supporting medium. We followed the methods described by Commuri and Keeling (2001) and Matsumoto et al. (1990). Using native polyacrylamide gels containing various concentrations of amylopectin and starch (both from maize, Sigma), the relative mobilities of wheat SSs and SBEs in monomeric and aggregated forms were measured at room temperature (23-25°C). Bovine serum albumen was used in control experiments and its mobility detected in gels by staining with Coomassie G-250 (see above). Paired-sample t-tests were used to compare the K d values of SBEII isoforms from the HMW and LMW fractions following electrophoresis in native gels containing corn starch.

Cross-linking
Amyloplasts were prepared for cross-linking experiments using buffers free of EDTA and DTT, and were lysed in gel filtration chromatography running buffer containing a protease inhibitor cocktail (see section above on plastid isolation).
Fractions of amyloplast lysates separated by size exclusion chromatography were immediately incubated with 1mM of the homobifunctional cross-linking reagent bis

Mass Spectrometry
In-gel digestion with trypsin and preparation of peptides for MS were as described previously (Tetlow et al., 2004b). Tandem electrospray mass spectra were recorded using Proteins were identified by correlation of uninterpreted tandem mass spectra to entries in SwissProt/TREMBL, using ProteinLynx Global Server (Version 1, Micromass). One missed cleavage per peptide was allowed, and an initial mass tolerance of 50 ppm was used in all searches. Cysteines were assumed to be carbamidomethylated, but other potential modifications were not considered in the first pass search. When this approach failed amino acid sequences were deduced manually from the charge state de-encrypted spectra (Wait et al., 2002), and were used as queries for searches using BLAST (Altschul et al., 1997) and FASTS (Mackey et al., 2002).

Protein Determination
The protein content of wheat endosperm whole cell extracts and plastid preparations was    Amyloplast lysates prepared from developing endosperm at 10-15 DAP were fractionated by gel filtration chromatography into two major peaks of SS and SBE activities (see  and amylopectin (C) in the gels at room temperature. Potential phosphorylation-dependent protein-protein interactions within amyloplasts of wheat endosperm are shown, as deduced from the coimmunoprecipitation experiments and cross-linking studies presented in this article. All the proteins shown in Figure 6 are present at the earliest stages of endosperm development used in the present study (6-9 DAP), but the various interactions shown are only detected later in development (10-15 DAP). It is assumed that all potential components of the complexes are present within the same cell-type. Broadly, three groups of potential interactions may be deduced from the experimental data; 1) homo-dimers of SBEII isoforms, 2) complexes formed between SS isoforms and SBEII homo-dimers, and 3) heterotrimeric complexes consisting of SSI, SSIIa, and one of the SBEII isoforms. In the present study there was no evidence for homo-or hetero-dimers of SSI/SSII forms, but we cannot rule out the possibility of complexes comprising SS dimers with an isoform of SBEII being present.